Lysine degradation-related metabolic remodeling in sarcopenia: Mendelian randomization prioritization and...

Previous metabolomic studies have implicated mitochondrial energy metabolism, lipid remodeling, and carnitine-related pathways in sarcopenia, but broader metabolic pathways linking these abnormalities remain unclear. We aimed to identify sarcopenia-related metabolic pathways by integrating metabolome-wide Mendelian Randomization (MR) with patient-derived serum and skeletal muscle multi-omics analysis.

Two-sample MR analyses were performed using genome-wide association study summary statistics for 1400 circulating metabolite-related traits and three sarcopenia-related traits, appendicular lean mass (ALM), handgrip strength, and walking speed. Candidate metabolites retained after MR sensitivity filtering were subjected to Kyoto Encyclopedia of Genes and Genomes (KEGG)-based pathway enrichment. The MR-prioritized pathway was then examined in an independent cohort of 40 postmenopausal women with or without sarcopenia using untargeted serum and skeletal muscle metabolomics. Patient-derived and the external public skeletal muscle transcriptomic datasets were analyzed to evaluate pathway-related transcriptomic patterns.

Lysine degradation was the only pathway meeting the predefined pathway-level FDR criterion in the ALM-negative direction. Lysine, carnitine, and N6,N6,N6-trimethyllysine contributed nominal negative MR associations to this pathway. Colocalization identified strong evidence of a shared causal signal for one carnitine locus (rs1171617; PP.H4 = 0.995). In the validation cohort, nine lysine degradation-related metabolites detected in serum and skeletal muscle were grouped according to three biochemical processes: free lysine oxidation, methyllysine turnover, and trimethyllysine-derived carnitine biosynthesis. In serum, dl-lysine was nominally higher (log₂FC = 0.796, P = 0.047) and dl-carnitine was nominally lower (log₂FC = -0.130, P = 0.047) in sarcopenia, while the trimethyllysine-derived carnitine biosynthesis score was significantly lower (adjusted P < 0.001). In skeletal muscle, γ-butyrobetaine (log₂FC = 1.178, P = 0.029) and N-acetyl-L-2-aminoadipic acid (log₂FC = 2.240, P = 0.035) were nominally higher in sarcopenia, whereas none of the three biochemical group scores differed significantly between groups. Patient-derived skeletal muscle transcriptomics were evaluated using predefined Reactome pathways for lysine catabolism and carnitine synthesis; the carnitine synthesis pathway score was lower in sarcopenia (P = 0.049), with significant downregulation of TMLHE (log₂FC = -0.411, adjusted P = 0.023) and DLD (log₂FC = -0.468, adjusted P = 0.004). In the external skeletal muscle dataset, TMLHE and DLD showed concordant decreases at nominal significance.

Lysine degradation was prioritized as an ALM-related metabolic pathway through exploratory MR-based pathway analysis. Multi-omics characterization indicated heterogeneous and compartment-specific remodeling rather than uniform alteration of the entire pathway, with the most coherent cross-omics pattern involving trimethyllysine-derived carnitine metabolism. These findings support lysine-related metabolism as a candidate metabolic framework for sarcopenia while warranting further mechanistic and longitudinal validation.